Pipelining Considerations - 2023.2 English

Versal Adaptive SoC Hardware, IP, and Platform Development Methodology Guide (UG1387)

Document ID
Release Date
2023.2 English

Another way to increase performance is to restructure long datapaths with several levels of logic and distribute them over multiple clock cycles. This method allows for a faster clock cycle and increased data throughput at the expense of latency and pipeline overhead logic management.

Because devices contain many registers, the additional registers and overhead logic are usually not an issue. However, the datapath spans multiple cycles, and you must make special considerations for the rest of the design to account for the added path latency.